Post on 17-Jan-2016
description
Frontier Models and Efficiency Measurement
Lab Session 3: Heterogeneity
William Greene
Stern School of Business
New York University
0 Introduction1 Efficiency Measurement2 Frontier Functions3 Stochastic Frontiers4 Production and Cost5 Heterogeneity6 Model Extensions7 Panel Data8 Applications
WHO Data
Heterogeneous Frontier Command
FRONTIER [; COST] ; LHS = the variable
; RHS = ONE, the variables, the additional variables ; EFF = the new variable $
ε(i) = v(i) +/- u(i)
Heterogeneous Frontier Model
FRONTIER ; LHS = LDALE ; RHS = ONE,LHEXP,LHEXP2,LEDUC,
VOICE,GEFF,LPOPDEN,TROPICS
; EFF = UI_WHO $
Heteroscedasticity
ui u i
vi v i
Either or both of
exp[ ]
(heteroscedasticity in inefficiency)
exp[ ]
(heteroscedasticity in random shift)
Variables may be the same or different.
w
w
Effects of Environmental Variable on Efficiency
? Frontier cost function with environmental? Variables, load factor, points served, stage? Length.FRONTIER ; Cost ; Lhs = lc ; Rhs = one,…, loadfctr,log(points),log(stage) $
? How does load factor affect efficiency?SIMULATE ; scenario: & loadfctr = .4(.025).95 ; plot $
Efficiency Estimates|Load Factor
Model Command for Heteroscedasticity
Heterogeneity in the Mean of u(i)
i
i i
2i
i
Mean of u depends on measured variables
u | U |
U ~ N[ , ] (may be heteroscedastic)
= i u
iz
Nonparametric Frontier
Frontier with local
Compute residuals then estimate and given the set of residuals from the nonparametric frontier
Compute efficiencies using residuals and and
Parametric vs. Nonparametric
Latent Class Model